An international team of scientists has revealed new clues about the neural mechanisms underlying memory formation and recognition in the developing human brain, according to a study published in Science Advances.
The ability to remember specific details of events and recognize familiar items is a critical cognitive function that depends on the formation and retrieval of memory traces, which can be tracked through the coordinated activity of neurons.
While scientists have been able to study the parts of the brain involved in memory formation and retrieval, how these processes are implemented by different parts of the brain has remained poorly understood, said Elizabeth Johnson, PhD, assistant professor of Medical Social Sciences, Pediatrics, and an assistant professor of Psychology in the Weinberg College of Arts and Sciences, who was co-senior author of the study.
In the study, investigators studied intracranial EEG recordings obtained directly from the brain in children, adolescents and young adults who previously had electrodes placed on their brain to monitor their epilepsy symptoms. The study participants were asked to view a visual scene and then later recognize what they saw.
“Our key finding is that recognition is supported by two distinct processes: reinstatement and transformation of memory traces initially formed during learning. We localized these processes to different regions of the brain and, in doing so, discovered how memory traces are reinstated and transformed to support successful recognition,” Johnson and co-authors said.
Reinstatement of items occurs in the temporal cortex, while transformation takes place in the parietal cortex, according to the study.
The findings also uncovered the roles of generalization and differentiation in neural networks, which allow the brain to adaptively encode and retrieve memories in visual and abstract formats. Scientists utilized a deep neural network to model the different formats of memory traces in the brain.
“Combined, our findings highlight the flexible and multifaceted nature of human memory traces and how they evolve during memory formation and retrieval,” Johnson said.
Moving forward, investigators will continue to explore memory traces using the datasets and neural network models developed in the study.
“We lay a foundation for further research on the formats and flexible nature of memory traces using advanced models and artificial intelligence,” Johnson and co-authors said. “Moreover, this unique dataset opens the door for a plethora of innovative studies on the development of memory and developmental neurophysiology broadly.”
Additional study authors included Nikolai Axmacher, PhD, professor at Ruhr University Bochum in Germany; Elias Rau, PhD, a student in the Axmacher lab; and Noa Ofen, PhD, professor at the University of Texas at Dallas.
The study was supported by grants from the European Research Council, the German Israeli Foundation, the German Research Foundation and the National Institutes of Health under grants R01MH107512, R00NS115918 and R01NS064033.
